Page |1
Manitoba Public Insurance
2017/18 GRA
MPI’s Investment Portfolio
Risk, Return and Good Practice
Prepared for Consumers’ Association of Canada, (Manitoba) Inc.
Submitted by the Public Interest Law Centre
Valter Viola
September 26, 2016
Page |2
CONTENTS
INTRODUCTION ............................................................................................................................................. 5
Terms of Reference ................................................................................................................................... 5
Rate Application Materials Reviewed ....................................................................................................... 5
Other Considerations ................................................................................................................................ 5
Organization of this Document ................................................................................................................. 6
Terminology .............................................................................................................................................. 7
AUTHOR BACKGROUND ................................................................................................................................ 7
Relevant Experience.................................................................................................................................. 7
CPP Investment Board ...................................................................................................................... 8
Ontario Teachers’ Pension Plan ........................................................................................................ 8
Pension Investment Association of Canada (PIAC) ........................................................................... 8
No P&C Experience ................................................................................................................................... 8
Teaching, Applied Research and Education .............................................................................................. 9
EXECUTIVE SUMMARY ................................................................................................................................ 10
Recommendations .................................................................................................................................. 10
I. OVERVIEW AND OVERSIGHT/GOVERNANCE ........................................................................................... 13
Principles that Informed Evidence and Questions .................................................................................. 13
Principle #1: The Valuation Basis Should Reflect the Specific Need ................................................... 13
Principle #2: Common Barriers to Excellence Should Be Acknowledged and Addressed .................. 13
Beliefs that Informed Evidence and Questions....................................................................................... 16
Belief #1: SUSTAINABILITY: The major stakeholder risk is that the current provisions will not be
sustainable in the future (recognizing that investment returns are one of many factors which will
contribute, positively or negatively, towards sustainability). ............................................................. 16
Belief #2 (MINIMUM RISK PORTFOLIO): Determining the Minimum Risk Portfolio is the first step
towards responsible long-term management of the portfolio........................................................... 16
Belief #3 (ADDITIONAL RISK): Taking additional risk beyond the Minimum Risk Portfolio should be
done only if the expected additional returns justify doing so. ........................................................... 17
Belief #4 (TOTAL PORTFOLIO): The additional risk to the Total Portfolio is the relevant risk to
consider if risk beyond the Minimum Risk Portfolio is taken. ............................................................ 17
Belief #5 (CONSTRAINTS): Constraints never increase expected risk-adjusted returns. ................... 19
II. MEASUREMENT ...................................................................................................................................... 20
Page |3
Background ............................................................................................................................................. 20
Different Users of Financial Information Have Different Needs ......................................................... 20
Useful Information Has Key Characteristics........................................................................................ 21
Accountants, Actuaries and Portfolio/Risk Managers are Guided by Different Principles................. 21
GAAP, ASOP and GARP are All Important, But They Answer Different Questions ............................. 22
Accountants .................................................................................................................................... 23
Actuaries ......................................................................................................................................... 23
Portfolio/Risk Managers ................................................................................................................. 23
Many Discount Rates, Each to Meet a Different Need ....................................................................... 23
Terminology Can Be Confusing, Given Different Bases for Measurement ......................................... 24
MPI’s Balance Sheet (GAAP) Needs “Adjustments” for Portfolio/Risk Management ........................ 25
Equities (18% of Assets) and HTM Bonds (25%) Require the Biggest Adjustments ........................... 27
Employee benefits (15%) and Unpaid Claims (70%) Require Adjustments Too ................................. 27
Q&A with MPI about Accounting and Portfolio/Risk Management Dependencies ............................... 28
Reconciling Market Values, Accounting Values and the Nature of their Risks ................................... 28
Accounting as a Potential “Tool” ........................................................................................................ 29
MPI Has Some Discretion re: Accounting, But It Is Not Clear What That Is ........................................ 29
Asset Accounting is Not Comparable Across the Portfolio ................................................................. 30
Asset Accounting: AFS Assets ............................................................................................................. 30
Asset Accounting: HTM Assets............................................................................................................ 31
Liability Accounting: “Remeasurement” of Employee Benefits (Excluded from Risk Metric) ............ 31
Greater Volatility in Comprehensive Income than Net Income.......................................................... 35
MPI’s Return/Risk Metric is a Less “Comprehensive” Accounting Metric .......................................... 36
MPI Elects to Use Net Income/Retained Earnings in Return/Risk Definition ..................................... 37
Short-Term vs Long-Term Tradeoffs ................................................................................................... 37
Comprehensive Income is Better than Net Income (for Risk Management), But Not Best ............... 37
Market Values are More Relevant than Accounting Values for Portfolio Decision-Making .............. 38
Market Risks Do Not Depend on How Assets and Liabilities are Measured by Accountants ............. 39
Three “Theoretical” Questions About the Interdependence of Accounting and Portfolio Design .... 40
1st Question: Changes in Asset Accounting ..................................................................................... 40
2nd Question: Changes in Liability Accounting ................................................................................ 40
3rd Question: Changes in Asset Mix ................................................................................................ 40
MPI’s Response to Questions 1, 2 and 3 ......................................................................................... 40
Page |4
III. MANAGEMENT....................................................................................................................................... 41
The Cost of Constraints ........................................................................................................................... 41
Reliance on Quantitative Models............................................................................................................ 41
Minimum/Maximum Constraints in Optimizations ................................................................................ 42
10% Minimum Weight to Canadian Equities to Retain a Meaningful Exposure to Home Markets ....... 42
Equity Risk Concentration ....................................................................................................................... 44
Duration .................................................................................................................................................. 46
Borrowing (Interest Rate Risk) in Real Estate ......................................................................................... 46
Commingling of Employee Benefit Plans ................................................................................................ 47
APPENDIX A: VALTER VIOLA’S CURRICULUM VITAE ................................................................................... 48
APPENDIX B: GLOSSARY OF TERMS ............................................................................................................ 51
Page |5
INTRODUCTION
Terms of Reference
I was retained by the Consumers’ Association of Canada Manitoba (“CAC”) to advise and assist on issues
related to the MPI investment portfolio (the “Portfolio”), including a good practice consideration of
issues related to risk versus return and prudent, reasonable and appropriate management of the
portfolio.
As stated in my terms of retainer, it is my duty to provide evidence that:



is fair, objective and non-partisan;
is related only to matters that are within my area of expertise; and
provides such additional assistance as the Public Utilities Board (“PUB”) may reasonably require
to determine an issue.
I understand that my duty in providing assistance and giving evidence is to help the PUB. This duty
overrides any obligation to CAC.
Rate Application Materials Reviewed
My review of the Rate Application focused on information related to the portfolio, including:



Asset-Liability Studies;
MPI’s Investment Policies; and
MPI’s Financial Statements and Accounting Principles.
Other Considerations
I also reviewed and considered the following other information:






MPI’s most recent annual report;
Information responses prepared by MPI and/or Aon to CAC questions relating to its investment
portfolio;
annual reports/websites of selected other funds;
selected information from Rate Applications related to prior years;
Report on Canadian Economic Statistics to 20131; and
Asset Mix Survey of Canadian defined benefit (“DB”) pension plans prepared by Pension
Association of Canada (“PIAC”).
To a lesser extent, I reviewed information about MPI’s capital adequacy (reserves). I also read the IRFRF
Technical Conference Materials and participated in the session by phone.
1
The latest report was not used. Instead, data from 2013 was used since this data would have been the data
available when the 2014 asset-liability study was completed by Aon.
Page |6
Organization of this Document
This document consists of these three sections:
I.
II.
III.
OVERVIEW AND OVERSIGHT/GOVERNANCE;
MEASUREMENT; and
MANAGEMENT.
This document takes into account three (3) inconvenient truths, and their consequences, which are
illustrated below.
TRUTH
Truth and
Consequence
HELPFUL
TOOLS
1.
2.
3.
We don’t know
very much
Beneficiaries/
owners ≠ managers
Correlations
are not perfect
Low signal/noise
(return/risk) ratio
Institutions manage $
for clients/beneficiaries
Risks don’t add
(1 + 1 ≠ 2)
We need to make
investment decisions
but
The interests of
stakeholders1
may not be aligned
because
The effect of an investment
on total portfolio risk depends
on the characteristics of
other assets in the portfolio
because
we don’t know
very much
beneficiaries/owners
are not the same
as managers
correlations are
not perfect



BELIEFS2
CAPITAL/REGULATORY
FRAMEWORK3
RISK FRAMEWORK3
Stakeholders include MPI, rate-payers and Manitoba Finance.
Beliefs support long-term portfolio design (e.g., asset mix).
Capital/Risk/Regulatory Frameworks provide a focus (clarity in mission) and process to overcome common barriers to excellence.
BELIEFS: In Section I (OVERVIEW AND OVERSIGHT/GOVERNANCE) my beliefs related to risk management
are summarized, along with a list of common challenges (“barriers to excellence”) among organizations
generally, and asset management firms in particular. I believe that two of the top three challenges
facing asset management firms relate to their focus and process, and the beliefs help to address these
challenges.
The third challenge relates to resources, which in MPI’s case, includes a portfolio that is closer to $2
billion than $200 billion. Simply put, practices that are appropriate and prudent for a $200 billion fund
may not be available to a $2 billion fund, though the beliefs and principles that are deemed to be
appropriate and prudent for both funds may be the same.
FRAMEWORKS FOR RISK/CAPITAL MANAGEMENT: The second and third truths (including their
consequences), inform Measurement (Section II) and Management (Section III) – i.e., long-term asset
mix decision-making.
Page |7
Terminology
Appendices include a glossary of terms to facilitate discussions, including the terms listed below. Not all
of these terms may appear in this document, but all terms were defined in either CAC (MPI) 1-67 or CAC
(MPI) 2-34.







Accounting reports
Actuarial reports
Asset risk
Assumptions
Available for sale (AFS)
Duration
Financial assets and liabilities at Fair
value through profit or loss (FVTPL)
Held to maturity (HTM)










Inherent risk
Investment risk
Liability from Basic Claims
Liability from pension plan
Liquidity risk
Market interest rate
Market reports
Market risk
Other comprehensive income (OCI)









Reinvestment rate risk
Risk
Risk (as defined in AssetLiability Study)
Risk-free rate
Risk profile
Surplus risk
Unrealized gains and losses
Valuation methodology
Value
Appendices include other information that is referenced in one or more sections.
AUTHOR BACKGROUND
Relevant Experience
With over 20 years of experience in the institutional fund management sector, I have 12 years of
combined executive, senior management and other experience at two of Canada’s largest institutional
investors:


~ $290 billion CPP Investment Board (2000 to 2005); and
~ $170 billion Ontario Teachers’ Pension Plan (1993 to 2000).
I have a decade of consulting experience (Holland Park Risk Management, since 2005) advising some of
North America’s largest institutional investors (boards, investment committees, executives and other
staff) on various aspects of investment risk management, measurement and governance. Clients have
also included Canadian pension funds with assets under management that are about the same as those
managed by MPI.
My specific expertise includes:



investment research, economics and risk management;
portfolio management; and
quantitative asset-liability modelling.
My curriculum vitae is in Appendix A.
Page |8
CPP Investment Board
I was the first executive officer responsible for research and risk management of CPP Investment Board
(“CPPIB”), the largest single purpose pool of capital in Canada with assets under management of ~ $290
billion today. As VP, Research and Risk Management I reported to the CEO/Chief Investment Officer over
a five year period and was the chief risk officer responsible for all aspects of investment risk
management. Research responsibilities covered a broad range of investment issues, including long-term
asset allocation, currency hedging and investment policies. As one of four investment executive officers,
I was involved in investment strategy specifically and the management of the organization generally.
Ontario Teachers’ Pension Plan
I held a variety of roles over a seven year career at Ontario Teachers’ Pension Plan (“Teachers’”), the
largest single-profession pension plan in Canada with ~ $170 billion of assets under management today.
Teachers’ is considered a global leader in pension/risk management.
As the first analyst in the Research and Economics Department, I developed Teachers’ first AssetLiability Model to support long-term asset allocation strategies (e.g., allocation between equities, bonds
and other asset classes). As a Director of Research, my responsibilities included conducting research to
support the introduction of new asset classes, currency hedging policies, as well as broad asset
allocation decisions on both a strategic (policy/long-term) and tactical (active/short-term) basis.
As a member of Teachers’ Investment Planning Committee, I participated in shorter-horizon
(tactical/active) decisions regarding asset allocation and currency hedging. As the first Portfolio Manager
of Teachers’ Tactical Asset Allocation (“TAA”) portfolio, I was responsible for managing one of the most
“active” programs at Teachers’2. I was also the Portfolio Manager of the real return bond (“RRB”)
portfolio3.
Pension Investment Association of Canada (PIAC)
I was also a member of the Investment Practices Committee of the Pension Investment Association of
Canada (“PIAC”).
No P&C Experience
I have not worked with property and casualty (“P&C”) insurers, but I have been an advisor to workers’
compensation funds in Canada and asset managers who manage assets for workers’ compensation
funds in both Canada and the United States.
2
Performance of the TAA portfolio exceeded value added targets and resulted in the maximum bonus multiplier
allowed under Teachers’ performance incentive system.
3
As RRB portfolio manager, I was part of a team that greatly increased the size of Teachers’ RRB portfolio. In the
year that I left Teachers’ to join CPPIB (2000), Teachers’ increased its exposure from $8.6 billion to $20.6 billion.
Page |9
Teaching, Applied Research and Education
I received the Chartered Accountant (CA) designation and taught accounting as well as
finance/investments at Wilfrid Laurier University and York University respectively before starting my
career in applied research and portfolio/risk management in the Research and Economics Department
at Teachers’, where I qualified as a Chartered Financial Analyst (“CFA”).
I am an MBA graduate from the Richard Ivey School of Business (Western University) and graduated
from the University of Toronto with distinction (BComm).
P a g e | 10
EXECUTIVE SUMMARY
This paper has been informed by:



principles that are generally accepted and relevant in the circumstances;
investment beliefs that, by definition have varying degrees of empirical support and theoretical
justification, but which I hold and consider to be reasonable and relevant in the circumstances;
and
facts from a variety of sources, including:
o MPI (e.g., the Rate Application, audited financial statements and responses to CAC
questions); and
o Aon (e.g., Asset-Liability Study and responses to CAC questions related to it).
Recommendations
Clarity of Accounting
Choices
MPI should clarify what flexibility it has regarding the accounting for assets and liabilities,
while remaining GAAP-compliant, and the factors it takes into account in electing to use
one method/assumption over others.
Adoption of More
Comparable
Accounting
Principles
In measuring its investment portfolio and liabilities, MPI should consider adopting
accounting principles, where GAAP allows MPI to make such elections, that reduce the
discrepancy between net income and comprehensive income (as these terms are
currently defined by MPI), to improve comparability across all assets as well as liabilities.
Comparability would be improved, for example, by accounting for more assets in a way
that is consistent with the treatment of financial assets and liabilities at fair value through
profit or loss (“FVTPL”).
AFS and HTM
Accounting
Unrealized gains and losses for AFS assets (~ 20% of assets), for example, are reported as
“other comprehensive income (OCI)” and are excluded from net income until realized,
making the net income recognition for unrealized gains on equities (~ 18% of assets)
inconsistent with FVTPL assets. The treatment of HTM Bonds (25%), recorded at
amortized cost, should also be re-considered.
Market valuations are generally more comparable, relevant, transparent, understandable
and subject to less potential bias than valuations in reports that are based on MPI’s
current accounting practices.
Pension Liability
Accounting
Reconsideration should also include the remeasurement of employee benefits (~ 15% of
liabilities and equities) which is considered OCI. The remeasurement of employee
benefits is large (given the long duration of pension liabilities), but OCI arising from
changing interest rates that impact the value of pension liabilities is not recognized
through transfers to net income under current practices.
Return/Risk
Definitions for Asset
Mix Decision
MPI should re-define return/risk used to inform its long-term asset mix decisions to be
based on valuations that reflect market values, rather than accounting ones (which may
be materially different).
At a minimum, net income should be replaced by comprehensive income in the
numerator (return) and retained earnings should be expanded to include accumulated
other comprehensive income (AOCI) in the denominator (risk).
In the long term, market returns and market risks will determine average long-term
premium rates, regardless of how assets and liabilities are accounted for under GAAP.
P a g e | 11
De-Linking Discount
Rates
For purposes of long-term asset allocation decision-making, MPI should consider
“breaking the link” (recursive) between liability valuations and the yield on some of its
assets. Economic theory suggests this approach is more appropriate.
Min/Max Asset Class
Constraints
The minimum/maximum and other constraints imposed on the portfolio (e.g., when
asset-liability studies are conducted) should be reviewed and relaxed, to avoid costly
constraints (lower risk-adjusted returns).
The rationale for imposing any such constraints should be made explicit.
Evolved Risk
Framework
An evolved risk framework should be considered to improve portfolio/risk measurement,
management and/or governance.
Explicit Risk
Management Goals
Among other things, the risk framework could include explicit goals related to market risk
management (as well as goals related to other types of risk if those require
enhancement). One goal might be to avoid “undue risk”, which is risk that is taken:


unknowingly, not having been identified (unaware); or
knowingly, but which:
o cannot be managed prudently, given current capacities (ineffective);
o exceeds risk tolerances (prohibited);
o is higher than it needs to be (inefficient); or
o is not understood (uninformed).
Another goal might be to get paid better/well for those risks that are desired, with
incentive systems that encourage desired behaviours that achieve desired outcomes and
controls that monitor compliance with limits that discourage/prevent undesired
behaviours and prevent undesired outcomes.
Minimum Risk
Portfolio
A minimum risk portfolio (for market risk) should be clearly defined. It should be aligned
with the interests of relevant stakeholders, with clarity regarding the short-term and
long-term factors that impact rate sustainability and other important outcomes.
Canadian Equities’
10% Minimum
Allocation
The appropriateness and prudence of having a 10% minimum weight to Canadian Equities
(“to retain a meaningful exposure to home markets") should be reconsidered, given the
different interests of different stakeholders (e.g., employees through the pension plan),
the concentrated nature of Canada’s equity market, and other such relevant
considerations.
No International
Equities
The appropriateness and prudence of having no exposure to International Equities should
be reconsidered, given the large size of non-US foreign markets, the return opportunities
that are potentially available from those missed opportunities and the effects of
increased international diversification on long-term market risks.
No Over-Reliance on
Quantitative
Modeling
MPI should be vigilant about its potential over-reliance on quantitative considerations,
given the high sensitivity of optimal asset allocations to seemingly small changes in capital
market assumptions (returns, volatilities and correlations) and the large number of
inputs.
Exclusion of Real
Return Bonds
The role that RRBs can play in effectively managing relevant risks should be discussed,
with consensus achieved regarding the effectiveness of RRBs from a risk management
perspective (i.e., independent of the cost of any “insurance” as measured by RRB yields
and their expected returns).
Effectiveness of
Duration Policy
The effectiveness of the duration policy should be reviewed, given the inherent risks of
changing real interest rates and unexpected inflation arising from MPI’s liabilities, and
exposure to changes in nominal interest rates in the MPI portfolio (i.e., nominal bonds
P a g e | 12
without inflation protection). More specifically, MPI should re-assess the effectiveness of
its duration-matching strategy since inflation (actual and/or expected) may differ from
current expectations.
Integration of Real
Estate/Infrastructure
Liabilities in
Duration
Management
MPI should consider the liabilities arising from all sources (i.e., including real estate and
infrastructure) in its interest rate risk management practices (duration), to be consistent
with its management of risks arising from insurance, pension and other liabilities.
Removal of 105%
Rule in Investment
Policies
MPI should remove from its Investment Policies the ability to request external managers
to realize gains (losses) (“105% Rule”), which MPI says “is no longer relevant”.
Pension Fund
The interests of all relevant stakeholders should inform decisions regarding both the
accounting for and management of the assets and liabilities related to the pension plan
and other employee benefits. A desirable outcome is to have greater clarity around the
appropriateness and prudence of maintaining different types of assets and liabilities
commingled in one fund.
The financial leverage assumptions used in Asset-Liability Studies that support long-term
asset mix decisions should be made consistent with the leverage actually used in the
portfolio, removing the ~ 4% difference related to real estate debt.
This would remove an ability by MPI to cause a manager to realize gains (losses) for the
sole purpose of having an impact on net income, without yielding any economic value,
reducing risk or otherwise conferring another benefit on MPI.
P a g e | 13
I. OVERVIEW AND OVERSIGHT/GOVERNANCE
Principles that Informed Evidence and Questions
Principle #1: The Valuation Basis Should Reflect the Specific Need
The basis for valuing assets and liabilities, and recognizing revenues, expenses, gains and losses, should
reflect the specific need (i.e., decision). Some needs are common to all stakeholders (e.g., assessment of
past performance) and can be met by the application of generally accepted accounting principles
(GAAP), with perhaps minor adjustments, while other needs are better met using other bases for
measurement that require much larger adjustments to GAAP.
I believe that long-term, asset mix decisions in particular requires the use of market valuations in
calculations of returns and risks rather than ones that are based on accounting valuations.
Principle #2: Common Barriers to Excellence Should Be Acknowledged and Addressed
Institutional investors face common challenges (“barriers to excellence”). For example, over 20 years
ago senior executives responsible for managing defined benefit pension plans cited the following three
largest barriers to excellence:
1. Poor process (almost 100%):
a. Structure;
b. Communication;
c. Inertia;
2. Inadequate resources (~ ½); and
3. Lack of focus or clear mission (almost ½).
This evidence and questions I posed were informed by the first and third common challenge noted
above. i.e., process (#1), including communication (i.e., accounting) and focus (#3), including how
return/risk is defined.
Given the nature and scope of my terms of reference, I include no observations regarding resources (#2
challenge).
The next page shows the source for these common barriers, including the rest of the “Top 10” list of
barriers, with “conservatism” being #4.
P a g e | 14
P a g e | 15
A Harvard Review article (2005) made similar observations (below) regarding organizational
performance generally (i.e., not just pension funds or other types of institutional investors), describing
“where the performance goes” and how average realized performance falls short (only 63%, per graph
below).
P a g e | 16
Beliefs that Informed Evidence and Questions
The first “inconvenient truth” is that we need to make investment decisions but we do not know very
much in the field of investments due to the low signal/noise (return/risk) ratio. As a result, we need to
rely on beliefs and I believe that it is important to be transparent about those beliefs. (MPI has included
some of its beliefs in its Investment Policies, and I list five of mine below.)
The five beliefs listed below were developed by a former colleague of mine at CPPIB, whom I consider to
be a global thought leader in investment management4.
Belief #1: SUSTAINABILITY: The major stakeholder risk is that the current provisions will not be
sustainable in the future (recognizing that investment returns are one of many factors which will
contribute, positively or negatively, towards sustainability).
This belief is important because “lack of focus or clear mission” was cited by almost ½ of surveyed
pension executives as being a large barrier to excellence (3rd largest challenge).
I believe that investment programs need to focus on the risk of long-term sustainability, and to develop
the key metrics (starting with risk) that clearly define the primary risk (i.e., primary risk metric, and time
horizon).
Belief #2 (MINIMUM RISK PORTFOLIO): Determining the Minimum Risk Portfolio is the first step
towards responsible long-term management of the portfolio.
This belief is important because “poor process” was cited by almost 100% of surveyed pension
executives as being the largest (#1) barrier to excellence. I believe that investment programs need to
have better processes for communicating the primary risk and the structures to support the
management of that primary risk (i.e., metrics that define return/risk tradeoffs, with a long-term
perspective).
I believe that MPI’s minimum risk portfolio (MRP) should include at least some long-duration real return
bonds (RRBs), given the nature of MPI’s liabilities (long term, with some inflation exposure). Note,
however, that Belief #2 simply supports the definition of the primary risk, but says nothing about
whether to buy any assets that make up the MRP (e.g., RRBs). The belief says nothing about how much
risk should be taken in relation to it. Appropriate and prudent answers to these follow-on questions
requires additional beliefs, starting with the next two beliefs below (taken together).
4
Donald Raymond led the development of these beliefs as Vice President, Public Markets at CPP Investment
Board. These beliefs were approved by the CPPIB Board of Directors, and reviewed by three other external
advisors who were considered to be thought leaders as well.
P a g e | 17
Belief #3 (ADDITIONAL RISK): Taking additional risk beyond the Minimum Risk Portfolio should be
done only if the expected additional returns justify doing so.
AND
Belief #4 (TOTAL PORTFOLIO): The additional risk to the Total Portfolio is the relevant risk to
consider if risk beyond the Minimum Risk Portfolio is taken.
Together, Beliefs #3 and #4 imply that the attractiveness of any asset class should be assessed, not in
isolation, but in the context of the other assets that are in the portfolio or contemplated to be in the
portfolio, and the MRP that is defined in Belief #1 and #2.
The rationale for taking a total portfolio approach stems from the “inconvenient truth”, mentioned
earlier, and re-stated below.
The effect of an investment on total portfolio risk depends on the characteristics of other assets
in the portfolio because correlations are not perfect.
In questions related to RRBs, MPI said:
“Real return bonds were excluded … because they were deemed to be expensive. Aon Hewitt’s …
assumptions showed real return bonds to have significant volatility and down side risk* with
modest returns relative to nominal bonds …. Also, page 17 of the Phase 1 report Aon concluded
that “RRBs are not a good inflation hedge”.” 5
* Emphasized by me, not MPI, to highlight a fundamental difference in beliefs.
When asked if MPI agreed with AON’s conclusion that RRBs are not a good inflation hedge, MPI said:
“The Corporation accepts Aon’s belief that there are other inflation hedging asset classes
available (i.e.: real estate and infrastructure) with greater expected returns … At the time of the
ALM study the real yields on RRBs were below 40 bps for 20 year terms and below 10 bps for 10
years and shorter terms. Real yields for the same terms are currently negative.”
Asked why Aon believes RRBs are not a good inflation hedge for MPI, Aon’s response was:
“RRBs are not a good inflation hedge for MPI for the following reasons:
1. The underlying inflation according to nominal and real return bonds do not match the
inflation used to value liabilities (which is based on a survey of Canadian banks);
2. RRBs suffer from a limited offering;
3. Supply and demand for RRBs have a large impact on the market value; and
4. Therefore, the economics of the inflation protection from RRBs do not match the financial
impact to MPI on a year by year basis.”6
* Emphasized by me, not Aon, and discussed on the next page.
5
Source: CAC (MPI) 1-77
6
Source: CAC (MPI) 2-39
P a g e | 18
My observations related to Aon’s four points are:
1. this is a problem with the method for valuing the liabilities (survey of Canadian banks),
which is not a market-based method, not a problem with the hedging properties of RRBs
against a market-valued set of liabilities;
2. while RRBs suffer from a limited offering (liquidity), larger funds have managed to
accumulate significant exposures (e.g., the average PIAC fund has 3%);
3. while supply and demand for RRBs may have a large impact on market value, presumably
this is a one-time market impact acquisition cost – a small price to pay if RRBs are
considered a buy-and-hold asset class, with little turnover; and
4. Aon appears to concede that RRBs do offer inflation protection (despite their earlier
comment to the contrary) but that RRBs do not match the financial impact to MPI on a
year by year basis. This last point represents the symptom of a bigger problem, which
relates to the next belief (constraints).
When asked if Aon could list one or two asset classes that offer better inflation hedges than RRBs for
MPI, and offer any evidence to support that belief, Aon said it could not, adding:
“There is no asset class that we know that can hedge the short term inflation risk … Over the
long term, where RRB’s are held to maturity, shorter term price sensitivity is less relevant and
inflation experienced over the period would result in higher cash flows and an inflation
hedge*. It is a commonly accepted belief … that higher inflation would gradually be reflected in
nominal bond yields, equity returns through higher profits, real estate through increased rents
and infrastructure, especially where regulated, through increased tariffs ...”
* Emphasized by me, not Aon, to note the tradeoff between shorter term price
sensitivity (less relevant according to Aon, with which I agree) and inflation
experienced over the period which would result in higher cash flows and an inflation
hedge.
P a g e | 19
Belief #5 (CONSTRAINTS): Constraints never increase expected risk-adjusted returns.
This belief is important because there appear to be two very binding constraints that may cause MPI to
have lower risk-adjusted returns.
These very binding constraints relate to:
1. HOW RISK IS DEFINED (i.e., with focus on the short term)
“The short-term volatility of the premium rate requirement is a primary concern”7; and
2. HOW ASSET CLASSES ARE CONSTRAINED
Minimum/maximum asset allocations in optimizations conducted as part of the most recent
Asset-Liability Study, for example, are overly restrictive.
The constrained optimization problem results in a “pre-determined” allocation to ≥ 80% of the
portfolio, leaving little room to optimize return/risk tradeoffs (i.e., allocating ≤ 20% of the
remaining assets). This is discussed further in another section.
7
Source: CAC (MPI) 1-73
P a g e | 20
II. MEASUREMENT
The Measurement Section consists of two parts:


Background; and
Q&A with MPI about Accounting and Portfolio/Risk Management Dependencies.
Background
Different Users of Financial Information Have Different Needs
There are a variety of stakeholders who have an interest in MPI, including:



MPI;
consumers, who are premium/rate-payers; and
Manitoba Finance.
Each stakeholder has common information needs regarding MPI’s financial affairs, as well as unique
ones that reflect their unique interest or relationship with MPI as well as their unique
roles/responsibilities.
To meet the information needs of such a varied group of stakeholders, it is not reasonable to expect one
set of reports to satisfy all needs of all stakeholders. That is why different kinds of reports are provided
by different types of professionals, including:



accountants;
actuaries; and
portfolio/risk managers.
These three information providers, and the bases for their reporting, are summarized below to facilitate
discussions and avoid potential confusion. While accountants, actuaries and portfolio/risk managers use
similar terms (e.g., “assets” and “liabilities”), the bases for assigning values to assets and/or liabilities
often varies and these differences can be material.
This is important because the appropriateness and prudence of relying on a particular set of metrics
(e.g., those prepared in accordance with GAAP) depends on the circumstances (i.e., the intended use of
the information). In MPI’s case, for example, MPI defines return/risk using accounting metrics (“net
income/retained earnings”), rather than metrics based on market values. As a portfolio/risk manager on
the one hand and former professional accountant on the other hand, I appreciate the usefulness of
different measurement bases for different purposes and the potential for confusion among stakeholders
that may arise.
Before reviewing the methods/assumptions used by each provider of information, and the
circumstances under which one method or source is more appropriate and prudent than others, it is
important to understand the characteristics of information that make it more useful in decision-making.
P a g e | 21
Useful Information Has Key Characteristics
The quality of information varies depending on many factors, and an assessment of information quality
starts with clarity and agreement on the specific need (decision support) and the relevance of the
information in meeting that need. Some of the characteristics of information that make it more useful
are listed below. To inform the development of GAAP, for example, the accounting profession suggests
that information is more useful if it is more:






relevant;
neutral/unbiased;
free from error;
comparable over time and across assets and liabilities;
verifiable; and
timely.
Tradeoffs exist, so choices need to be made. For example, information that is most relevant (e.g., about
an uncertain future) may be less reliable compared to the past (no uncertainty, since it happened).
Accountants, Actuaries and Portfolio/Risk Managers are Guided by Different Principles
Each type of professional is guided by principles and standards that are considered appropriate and
prudent in their profession.
Accountants are guided by generally accepted accounting principles (“GAAP”).
Actuaries are guided by actuarial standards of practice (“ASOP”).
Portfolio/risk managers are guided by generally accepted risk principles (“GARP”).
P a g e | 22
GAAP, ASOP and GARP are All Important, But They Answer Different Questions
GAAP, ASOP and GARP have similarities, but also many differences. As the table below illustrates,
reports prepared by accountants, actuaries and portfolio/risk managers differ in at least three respects:



time period (past, present and future);
basis for valuation and conservatism (e.g., historic/amortized cost, market value, etc.); and
single valuation, compared to a range.
Role
Accountant
Relative8 Focus
(Scale from 1 to 10)
Principles/
Standards
Perspective
Past
Present
Future
5
4
1
●●●●●
●●●●
●
What
happened?
Today and Past
(last 2 years)
GAAP
Questions

Decisions

Performance
Assessment
0
Actuary
7
●●●
●●●●●●●
Today and
Future
(decades)
ASOP
0
Portfolio/
Risk
Manager
3
GARP
2
8
●●
●●●●●●●●
Today and
Future
(decades)
What is
expected to
happen?

Funding
(Premiums/
Rates)
What could
happen
(good and
bad)?

Return/Risk
Profile
(Asset Mix)
8
9
Key Outputs
(Answers)
GAAPcompliant
Financial
Statements
(balance sheet,
income
statement)
ASOP compliant RateSetting
Processes,
Analyses and
Valuations
Efficient
Frontiers
(Expected
Returns/Risks9)
Numbers are indicative and designed to illustrate different time perspectives.
The author believes strongly that return/risk metrics should be based on the market value (as distinct from
accounting value) of assets and liabilities.
P a g e | 23
The following observations are intended to distinguish between the perspectives of accountants,
actuaries and portfolio/risk managers to facilitate discussions. The observations are not meant to be a
complete list of how the three professionals satisfy the requirements of their profession.
Accountants
Accountants prepare accounting balance sheets (statements of financial position) at a point in time (the
present), and income/cash flow statements about the recent past (prior two years).
While some values related to “present valuations” involve expected future cash flows, most reporting
relates to the present financial condition as well as the results of operating, investing and financing
activities from the recent past.
The present valuation of future cash flows, for the purpose of assigning a value to assets and/or
liabilities, is prescribed under GAAP with GAAP-compliant practices that vary between different types of
organizations (public sector vs private sector), different jurisdictions (Canada vs United States) and
different sectors (insurance companies, pension plans, etc.).
Actuaries
Actuaries prepare actuarial valuations at a point in time (the present), and project income/cash flows
into the future (decades). Their actuarial reports help to inform decisions about the sufficiency of assets
in relation to the liabilities at a point in time and possibly the sustainability (i.e., rate adequacy) of an
insurance plan or pension plan, as the case may be, over the longer term.
Portfolio/Risk Managers
Portfolio/risk managers prepare valuations at a point in time (the present) and assessments of the
market value at risk arising from potential future scenarios that vary in terms of their
likelihood/probability as well as severity/attractiveness (adverse vs favourable related to assets,
liabilities or both).
Unlike accountants, portfolio/risk managers measure opportunity costs, as well as actual or projected
costs. For example, a portfolio manager may consider the return of a benchmark portfolio (e.g., S&P
500) and a risk manager may consider risks in relation to that same benchmark portfolio (“tracking
error” or active risk) as well as in relation to a minimum risk portfolio that serves as a proxy for
underlying liabilities (“surplus at risk” or funding risk).
Many Discount Rates, Each to Meet a Different Need
For accounting purposes, MPI uses a discount rate for valuing liabilities that depends on yields related to
MPI’s assets. For actuarial purposes, MPI uses a discount rate that reflects the principles and standard
for that profession. For portfolio/risk management purposes, discount rates should reflect market
opportunities, and these rates may differ significantly from both accounting and actuarial discount rates.
P a g e | 24
Terminology Can Be Confusing, Given Different Bases for Measurement
The next page starts with MPI’s balance sheet, prepared in accordance with GAAP, to facilitate
discussions and to contrast GAAP (accounting perspective) on the one hand with GARP (portfolio/risk
management perspective) on the other hand.
A glossary of terms is included in Appendix B.
For example, “comprehensive income” is the sum of “net income” and “other comprehensive
income (OCI)”.
These definitions are important because MPI uses accounting definitions for return/risk analyses that
support long-term asset mix decisions, and these are not the same as market-based definitions (which
the author believes are more relevant to support long-term asset mix decisions).
The next page illustrates the “adjustments” that a portfolio/risk manager might make to MPI’s
accounting values, to meet his/her asset allocation/risk management needs, using information from
MPI’s financial statements that MPI confirmed (in a later section) was correctly interpreted by me.
P a g e | 25
MPI’s Balance Sheet (GAAP) Needs “Adjustments” for Portfolio/Risk Management
The table below shows accounting values as a percentage (%) of the total portfolio from MPI’s recent
financial statements, including three (3) types of assets – each with a different accounting basis.
Three colours highlight the extent to which the
bases for accounting differ from market
valuations (red = larger potential difference,
yellow = more modest differences, green =
immaterial differences, if any).
To appreciate differences between accounting
and market valuations, the table below groups
assets vertically based on MPI’s accounting
(FVTPL, AFS, HTM) and horizontally based on
market volatility10.
Assets are sorted on an accounting basis, with
asset groups characterized as having smaller
differences between market values and
accounting values (green) at the top and larger
differences (red) below. Assets are sorted on a
market volatility basis with the least volatile
assets on the left. Liabilities are grouped
similarly, noting that the duration of employee
benefits (including the pension) is higher than
for unpaid claims (i.e., greater volatility).
Accounting Asset (Liability)
Accounting
Asset Group
FVTPL
Difference Between Market
and Accounting Value
Financial assets
and liabilities at
fair value through
profit or loss
Immaterial,
if any
AFS
Available for sale
More modest
HTM
Held to maturity
+ non-financial
Largest
potentially
A
B
GAAP
Reclassify
A+B
Reclassified
GAAP
3
13
(4)
42
2
18
25
1
100
Lowest
Cash
Market Volatilities
Real
InfraBonds
Estate
structure
3
13
(4)
42
Infrastructure
33
*
Real Estate – Pool
9
4*
FVTPL
Liabilities – Real Estate
(4)*
11
FVTPL Bonds
42
Cash12
2
2
AFS
Equities
18
HTM Bonds
25
0
25
HTM13
Real Estate – Property
1
1
Total Portfolio
100
0
2
63
14
3
Non-Portfolio Liabilities
Unpaid Claims
(70)
(70)
(70)
Employee Benefits
(15)
(15)
Total Equity
15
15
* Liabilities arising from real estate are reclassified for presentation purposes to “bonds” (by 4%)
to be more consistent with assumptions used in Asset-Liability Studies. For simplicity, a similar
adjustment could be made, but has not been, for liabilities arising from infrastructure given its smaller
size and consistent treatment between asset-liability studies and implementation in practice.
10
The order of market volatilities, from left to right, reflects the assumptions used in Aon’s Asset-Liability Study.
Includes Federal, Other Manitoba, Other Provinces and Corporates.
12
Includes other assets (not material at ~ 0.1%).
13
Accounting is assumed to be HTM. If this is not correct, the discrepancy is immaterial (only a 1% allocation).
11
Highest
Equities
18
18
(15)
P a g e | 26
The biggest differences between market values and accounting values are illustrated below, with assets
shown in the top panel and liabilities in the bottom.
The X axis shows market
volatility while the Y axis
shows accounting value
volatility.
If there were no differences
between the bases for
measuring assets/liabilities
for accounting purposes on
the one hand and
portfolio/risk management
purposes (market value
based) on the other hand,
assets and liabilities would
plot along a straight 45
degree line (Y = X; i.e., no
difference).
Assets/liabilities plotting below the line have market valuations that are more likely to deviate from
their accounting valuations, on average. The larger the vertical gap from the line, the larger the
difference is likely to be between accounting and market-based metrics. Portfolio/risk managers,
including the author, may make adjustments to accounting values in their analyses to reflect these
differences.
In MPI’s case, notable differences from the above graphs arise for several items.
P a g e | 27
Equities (18% of Assets) and HTM Bonds (25%) Require the Biggest Adjustments
Equities (18%) are the most volatile assets, but the material component of equity volatility is recognized
in net income only when gains/losses are realized through sales (i.e., unrealized gains/losses are
included as OCI, not net income). HTM bonds (25%) are recorded at amortized cost for accounting
purposes, which differs from market value.
Employee benefits (15%) and Unpaid Claims (70%) Require Adjustments Too
Employee benefits (15%) are the most volatile liabilities, given their long duration, but remeasurement
of Employee Future Benefits Gains/Losses do not get recognized in net income (only OCI). Liabilities
related to Unpaid Claims are valued for accounting purposes using a discount rate that is based on
certain assets in MPI’s portfolio, rather than a market rate that would be used by a portfolio/risk
manager.
P a g e | 28
Q&A with MPI about Accounting and Portfolio/Risk Management Dependencies
This section reviews some of the questions that CAC posed and the responses from MPI and/or Aon. The
questions were designed to better understand the dependencies, or lack thereof, between portfolio/risk
management on the one hand (e.g., asset mix choice) and the basis for measurements (values of assets
and liabilities, return/risk) on the other hand.
Reconciling Market Values, Accounting Values and the Nature of their Risks
The table below shows accounting values as a percentage (%) of the total portfolio using accounting
values from a recent MPI balance sheet, consistent with those shown earlier.
MPI confirmed that our understanding of the accounting, as depicted above, was correct by saying:
“your understanding is correct for the liabilities; however, please note that it is only the
remeasurement of the employee future benefits that is put to Other Comprehensive Income
(OCI), other items relating to employee future benefits is expensed to net income.”14
14
Source: MPI’s response to CAC (MPI) 1-71
P a g e | 29
I interpret this to mean that the (less uncertain/volatile) accrual related to employee future benefits is
reflected in net income (and therefore included in the risk metric, as defined in the Asset-Liability
Study), but that the (more uncertain/volatile) component arising from changing interest rates is not
reflected in MPI’s risk metric. (As noted in the notes to MPI’s financial statements, the duration or
interest rate sensitivity that results in “remeasurement” is high at 16 years – a topic discussed again in
more detail in another section).
MPI did not comment on the “asset” side of the depiction above, so I assume that MPI is in general
agreement with both the nature and size of the components of the assets, using the accounting-based
valuations from MPI’s financial statements.
Accounting as a Potential “Tool”
MPI was asked if one “general” tool that is either being used by MPI now, or could be used by MPI in the
future, is the “choice” of metrics (“basis for measurement”) for valuing assets, liabilities and income
recognition. MPI said:
“The “choice in metrics” for “valuing assets, liabilities and income recognition” is determined by
… IFRS ... These items are not at … MPI … discretion.”15
MPI Has Some Discretion re: Accounting, But It Is Not Clear What That Is
MPI was asked what flexibility it has in choosing accounting valuation methods for assets and liabilities
and what their implications for net income and OCI are. MPI said:
“… MPI … is required to adhere to … IFRS ... Where there are choices within an IFRS standard, the
Corporation evaluates the options and determines which is most appropriate for the
Corporation.”16
MPI also said:
“certain interest rate movement can be elected to be in either Comprehensive Income or Net
Income.” 17
15
Source: MPI's response to CAC (MPI) 2-34
Source: MPI’s response to CAC (MPI) 1-72
17
Source: MPI's response to CAC (MPI) 2-34 f)
16
P a g e | 30
Asset Accounting is Not Comparable Across the Portfolio
The inconsistent valuation and treatment of unrealized gains and losses across asset classes is an
undesirable characteristic of MPI’s accounting metrics (lower transparency, less comparability), from the
perspective of an external stakeholder (e.g., CAC). Some of this inconsistency may be due to the
required application of GAAP in order to be GAAP-compliant while some of it may arise from MPI’s
election(s) to choose one method where GAAP provides such choice(s).
Asset Accounting: AFS Assets
Net income under GAAP is a metric that is influenced (positively or negatively) by both:


the underlying transactions (“reality”); and
how transactions are reported (i.e., accounting policies, which change over time, vary by
industry/sector and region, and may provide some discretion or election to choose between two
or more practices that are generally accepted and GAAP-compliant).
The adoption by MPI of AFS accounting for equities (~ 18% of assets), for example, implies that only
realized gains will impact net income while unrealized gains will impact OCI. This means that net income
depends on equity “turnover”, which is controllable directly by the manager and indirectly by MPI.
(Comprehensive income would not change as a result of turnover alone, except for the related
transaction costs.) The potential for MPI to impact net income in this way, without adding real value as a
result of the transaction, is a governance feature that can be improved upon.
The Investment Policy Statement says:
“The Corporation may … request external equity investment counsels to realize capital gains that
have accumulated …. The ability to realize gains is subject to the ratio of the market value to
book value of the total of both the Canadian and U.S. equity portfolios (in their native currencies)
exceeding 105%. The market value to book value ratio must not fall below 105% as a result of
the gains taking process.”18
When asked what circumstances would cause MPI to request managers to realize gains that had accrued
and about the significance of the “105% rule”, MPI replied:
“The Corporation has not requested equity managers to realize capital gains for several years
and does not foresee any reasons to request the managers to realize capital gains in the future.
The 105% rule was added to the Investment Policy Statement (IPS) during the 2005 IPS Review.
The rule was established to provide a buffer from equity market volatility. This rule is no longer
relevant to the management of the investment portfolio.”
18
Source: CAC (MPI) 1-89
P a g e | 31
Asset Accounting: HTM Assets
Because HTM assets are carried at amortized cost in MPI’s financial statements19, some of the market
risks for a significant portion of the portfolio (~ ¼ of MPI’s assets) are not captured in the risk definition
used by MPI in long-term asset-liability studies (i.e., the risk definition is an accounting one, which
“under-reports” the true market volatility for HTM assets, among other assets and liabilities, as noted
elsewhere in this document).
Liability Accounting: “Remeasurement” of Employee Benefits (Excluded from Risk Metric)
MPI said:
“it is only the remeasurement of the employee future benefits that is put to … OCI … other items
relating to employee future benefits is expensed to net income”20.
When asked if remeasurement of the employee future benefits that is put to OCI is more volatile than
the “other items relating to employee future benefits”, MPI said:
“The response to this question requires a definition of ‘volatility’. The split between
remeasurement of employee future benefits and other items relating to employee future benefits
is based on accounting standards. Both are actuarially determined and fluctuate year over year.”
MPI was asked if it agreed that “remeasurement of employee future benefits (as defined in Note 21 of
MPI’s financial statements”) will never impact “net income” nor “retained earnings” as those terms are
used in the definition of “return” and “risk” in the Asset-Liability Study. MPI’s response is below.
“The Corporation does not strictly agree …, as accounting standards related to the
remeasurement of employee future benefits could change in the future. The Corporation …
follows all relevant accounting guidelines (IAS 19) …, but has no control over prospective changes
in those guidelines. Accordingly the Corporation cannot agree … that remeasurement of
employee future benefits will never impact net income.”
In response to i) AON offers the following:
We agree that the inherent economic or market risk arising from employee future benefits
(reflected in the re-measurement of employee future benefits) is not reflected in the return/risk
metrics in the Asset-Liability Study.
There is a tenuous link between retained earnings and re-measurement of employee future
benefits. Re-measurement … impacts … AOCI … AOCI is considered in the calculation of surplus
distributions/special contributions. Surplus distributions/special contributions impact retained
earnings. However, the impact is likely very small.”
Re-measurement of employee future benefits does not impact net income.
The impact of changes in the present value of pension liabilities is modeled and flows into the
AOCI.”
19
20
Source: MPI’s audited financial statements (Feb 2015)
Source: MPI’s response to CAC (MPI) 1-71
P a g e | 32
Based on these responses, I conclude that the inherent economic or market risk arising from employee
future benefits (reflected in the re-measurement of employee future benefits), which I believe to be
material, is not reflected in the return/risk metrics in the Asset-Liability Study and therefore not
appropriately considered in asset mix decisions. As noted in MPI’s financial statements “the weighted
average duration of the defined benefit obligation is 16.29 years”21. This means that a 1% change in the
discount rate used to value the liabilities would cause the value of the liability to change by
approximately 16% in the opposite direction.
From MPI’s financial statements (below, enlarged on the next two pages), the 0.45% increase in
discount rate (to 4.05%) resulted in a remeasurement gain of ~ $27 million last year, consistent with
MPI’s sensitivity analysis (next page) which shows an expected $51 million gain22 for a 1% increase (~
double the 0.45% actually reflected).
21
22
Source: MPI’s audited financial statements
$51 million gain ~ 16 year duration x 1% change x $312 million average value of pension benefits in the year).
P a g e | 33
P a g e | 34
Note 16 is continued below.
P a g e | 35
Greater Volatility in Comprehensive Income than Net Income
MPI was asked if “comprehensive income” (i.e., including OCI) is more volatile than “net income” (i.e.,
excluding OCI). MPI said:
“the answer depends on the definition of volatility and nature of volatility being assumed. If
there is more potential for fluctuation due to valuation changes in equities and employee future
benefits, then agreed.”
A portfolio/risk manager’s definition of volatility would be based on the market valuations of assets and
liabilities (as distinct from accounting valuations), implying that comprehensive income is more volatile
than net income as a result of investments in equities and employee benefits (mainly the pension plan’s
liabilities).
This greater volatility in comprehensive income, compared to net income, arises because
comprehensive income includes OCI, but net income does not, for:


unrealized gains and losses from changes in the market value of equities (AFS assets, which
includes equities), noting that equities have higher market volatility compared to other assets;
and
unrealized gains and losses from changing interest rates that impact the value of pension
liabilities, noting that the long duration (16 years, per MPI’s financial statements) of pension
liabilities makes these liabilities very sensitive to changes in interest rates.
MPI was asked if “comprehensive income” is closer to a market-based definition of “income” than is
“net income”, as these terms are currently defined. MPI’s response is below.
“This question is theoretical in nature and the response requires a definition of ‘market based’.
Comprehensive Income and Net Income are determined based on adherence to Accounting
Standards and available elections under the accounting standards. Certain interest rate
movement can be elected to be in either Comprehensive Income or Net Income. Other actuarial
valuations based on interest rates are required to be in comprehensive income based on
accounting standards.”
I do not agree with MPI regarding the theoretical nature of the question. I believe that comprehensive
income, as currently defined, is more volatile than net income, as illustrated in MPI’s most recent
audited financial statements below23.
23
Source: MPI’s audited financial statements (Feb 2016)
P a g e | 36
OCI is the difference between Net Income and Comprehensive Income, which consists of:


“items that will not be reclassified to income”
– i.e., Remeasurement of Employee Future Benefits Gain (Loss); and
“items that will be reclassified to income”
– i.e., Unrealized gains (losses) on Available for Sale assets (inclusion + reversal or
reclassification to AFS when realized as a result of a sale/turnover).
When asked if MPI agreed that “total equity” (i.e., including AOCI) has more volatility than “retained
earnings” (i.e., excluding AOCI), MPI said “the answer depends on the definition of volatility and nature
of volatility being assumed. If there is more potential for fluctuation due to valuation changes in equities
and employee future benefits, then agreed.”24
MPI’s Return/Risk Metric is a Less “Comprehensive” Accounting Metric
“Return” and “risk” are measured in Aon’s Asset-Liability Study on an “accounting” basis:


“net income”, on the return (Y) axis; and
“average annual volatility of retained earnings”, on the risk (X) axis.
MPI confirmed that:
“’return’ and ‘risk’ metrics in the Asset-Liability Study reflect the accounting definitions for ‘net
income’ and ‘retained earnings’ (i.e., they exclude … OCI … and … AOCI) in ‘return’ and ‘risk’
respectively)”.25
24
25
Source: MPI's response to CAC (MPI) 2-34
Source: CAC (MPI) 1-73
P a g e | 37
MPI Elects to Use Net Income/Retained Earnings in Return/Risk Definition
MPI said that it is not required, but elects to use “net income” and/or “retained earnings” (excluding
AOCI), rather than “comprehensive income” (including OCI) or “total equity” (including AOCI),
respectively, in “how return and/or risk are defined in Asset-Liability Studies”.26
The decision to use net income/retained earnings in Asset-Liability Studies has the effect of underreporting the volatility in Asset-Liability Studies compared to the underlying volatility that is actually
inherent in those assets and liabilities on a market value basis.
Short-Term vs Long-Term Tradeoffs
MPI was asked if there is a trade-off between short-term rate stability on the one hand and long-term
accuracy of market risk assessments that arises from using the same basis for measurement for both of
these two purposes. MPI responded:
“As stated on the PUB (MPI) 1-20 (a) Attachment B Phase II - Part A ALM Study page 5, MPI’s
Basic compulsory program is required to break even rather than to target profits. The objective
of the ALM study was that “the short-term volatility of the premium rate requirement is a
primary concern”. Therefore, the benefits of long-term accuracy of market return and risk
through comprehensive net income are not as important compared to reducing the short-term
volatility of the premium rate requirement.”27
MPI said that “short term volatility of the premium rate” is a primary concern, which was echoed by Aon
in the Asset-Liability Study, by saying Aon perceived that “MPI has a low risk tolerance resulting from the
mandate to break even instead of targeting profit, the extensive process to change targeted levels of
reserve and the lack of control of MPI over premium rates.”
MPI said that “for these organizations in year performance is less critical as they do not set annual rates
as does MPI. As a result, these organizations are able to take more investment risk and have a much
higher allocation to equities than MPI (an average of 50.8% vs. 18.5% in 2015). The average allocation to
fixed income by these organizations is 31.7%, while MPI had 66.4% in fixed income.”28
Comprehensive Income is Better than Net Income (for Risk Management), But Not Best
When asked if MPI believes that comprehensive income is a better or worse metric than net income, as
these terms are currently defined, MPI said:
“MPI would not define one measure as better than another. The two measures show different
views of the corporation’s activities which does not necessitate superiority of one over the
other. One may be more suited to an analysis than the other, depending on what is being
analyzed.” 29
26
Source: MPI’s response to CAC (MPI) 2-34(c)(ii)
Source: MPI’s response to CAC (MPI) 2-36(b)(ii)
28
Source: Volume II – Investment Income, page 25
29
Source: MPI’s response to CAC (MPI) 2-34(n)
27
P a g e | 38
I believe that the better metric depends on the principles/beliefs used to inform decision-making and
the purpose for which the metrics are used. For the following reasons, I believe comprehensive income
is better than net income, but still not the best, for the purpose of long-term asset allocation decisionmaking.




Comprehensive income is more relevant than net income, reflecting OCI components
that bring it closer to a market value basis;
Comprehensive income is more “comprehensive” or complete than net income,
because it includes unrealized gains and losses on AFS assets (20% of assets, including
equities which are more volatile than other assets) and employee future benefits
gains/losses (which have a high sensitivity to interest rates), while net income does not;
Comprehensive income is more neutral because it is not impacted (as net income is) by
portfolio turnover; and
Comprehensive income is more consistent/comparable and understandable, by
including more (but not all) assets and liabilities on the same valuation basis – market
value, as distinct from an accounting basis that currently includes some but not all of
unrealized gains/losses in income.
Any increased reliance on comprehensive income would likely involve higher costs (e.g., market
valuations), but these costs are likely to be small compared to the benefits which would flow from its
use (risk management generally, but long-term asset allocation specifically).
Market Values are More Relevant than Accounting Values for Portfolio Decision-Making
GAAP is designed to meet the common information needs of a variety of users who have different
needs. GAAP measures of assets and liabilities may be appropriate for some purposes but are less
relevant for others despite being “generally accepted” for accounting purposes.
Also, accounting (and actuarial) principles may differ from one jurisdiction to another, and they may
change over time. These differences and changes, however, are unrelated to the underlying factors that
create the inherent market risks related to the assets and liabilities. Simply put, measuring something
differently does not change its characteristics.
Portfolio/risk managers generally agree that market values (rather than accounting values) are more
relevant for informing investment/risk management decisions. This explains why portfolio/risk
managers make various “adjustments” to net income or asset/liability valuations in making decisions
about long-term portfolio designs (e.g., asset mix).
P a g e | 39
Market Risks Do Not Depend on How Assets and Liabilities are Measured by Accountants
MPI was asked if the market risk (e.g., volatility or other appropriate market risk metric) of MPI’s
liabilities depends on the principles, valuation methodologies, assumptions and/or values developed by
actuaries.
MPI responded by referring to its earlier response (CAC (MPI) 1-72 (f)), where MPI was asked if the
inherent market risks arising from the liabilities (viewed in isolation, not in relation to assets) can only be
changed by changing the nature of the cash flows that underlie the liabilities (e.g., by changing their
degree of indexation to inflation or other such basis), and not through the adoption of different
“valuation” methodologies (including “smoothing”).
In 1-72 f), MPI said:
“The Corporation agrees that inherent market risks arising from liabilities can be changed by
changing the nature of cash flows underlying the liabilities. However, such change can also occur
by the adoption of a different ‘valuation’ methodology. E.g. If the ‘valuation’ methodology
required the use of a risk-free interest rate to discount the liabilities, then the inherent market
risk would be reduced. However, the tradeoff is that MPI would have to hold a higher amount of
liabilities. As stated in the responses above, the Corporation’s current ‘valuation’ methodology is
in adherence to IFRS and is in accordance with current accepted actuarial practice”.
MPI’s responses to CAC (MPI) 1-72 e) appears to contradict its response to 1-72 f) regarding the
dependency of market risks and their basis for measurement. In 1-72 e), MPI agreed that the “’actual’
market risks arising from MPI’s liabilities do not depend on how those liabilities are valued (by either
accountants or actuaries)”.
The underlying or inherent risks related to assets and liabilities do not depend on how they are
accounted for, in the same way that using a Celsius thermometer (rather than Fahrenheit) leaves the
temperature unchanged despite having different readings on the two scales.
MPI’s long-term portfolio decisions (e.g., asset mix) are influenced by return/risk accounting-based
metrics, and these accounting metrics may differ materially from those that are “market-based”.
P a g e | 40
Three “Theoretical” Questions About the Interdependence of Accounting and Portfolio Design
MPI was asked if it agreed with the following three statements, the first two relating to potential
changes in accounting for asset and liabilities, and the third to the impact such accounting changes
might have (e.g., lead to more, less or about the same allocation to equities, as compared to bonds).
1st Question: Changes in Asset Accounting
 Assuming that GAAP were to change and require MPI to use market interest rates/prices for
valuing assets (i.e., 100% of unrealized gains and losses would impact net income):
 net income for accounting purposes would become more volatile;
 rates (i.e., premiums) would become more volatile (explain why; e.g., through the effects on
both net income and the fact that the rate-setting formula is based on net income); and
 the inherent (market) risk, as defined above, of the assets that would be subject to the
revised accounting treatment would remain unchanged. If not, why not?
2nd Question: Changes in Liability Accounting
MPI was asked if it agreed with the following statements, assuming that GAAP were to change and
require MPI to use market interest rates/prices for valuing liabilities:




net income for accounting purposes would become more volatile;
rates (i.e., premiums) would become more volatile. If so, please explain why;
the inherent (market) risk, as defined above, of the liabilities that would be subject to the
revised accounting treatment would remain unchanged; and
if not, please explain why not?
3rd Question: Changes in Asset Mix
MPI was asked if the changes to GAAP related to either the assets, liabilities or both described above
had been in place when the latest Asset-Liability Study was completed, what impact would the change(s)
likely have had (e.g., likely lead to more, less or about the same allocation to equities, as compared to
bonds)? Please explain briefly.
MPI’s Response to Questions 1, 2 and 3
MPI’s response to all these questions is below.
“This question is theoretical in nature and the request is to comment on potential currently
undrafted accounting changes that may or may not occur. However, please note that MPI
currently uses market interest rates and prices for valuing assets and liabilities where required
under current accounting standards.”
My question may be theoretical, but I believe it is an important one to ask.
Also, MPI does have some discretion regarding its accounting choices (as discussed elsewhere).
An article, funded by the Rotman International Centre for Pension Management (ICPM), that illustrates
how and why accounting matters (for better or worse) is included as a reference in the footnote
below30.
30
See Pension Fund Asset Allocation and Liability Discount Rates: Camouflage and Reckless Risk Taking by U.S.
P a g e | 41
III. MANAGEMENT
The Cost of Constraints
As noted in the beliefs section, “constraints never increase expected risk-adjusted returns”.
Nevertheless, there may be good reasons for imposing constraints, including some of the following
perhaps:






Legislation or regulation (e.g., maximum foreign property);
Market risk management (e.g., avoid concentration by setting a maximum);
Liquidity risk management (e.g., set a maximum for illiquid assets as a group and/or at the asset
class level for real estate, infrastructure and private equity);
Return expectations (e.g., maximum for low-yielding assets);
Insufficient internal/external asset management capabilities; and
Concerns about the accuracy of modelling methodologies and/or assumptions, and the (widelyknown) sensitivity of optimization results to assumptions re: returns, risks, correlations, etc.
Reliance on Quantitative Models
On this last point above, it important to note that:


optimal solutions from quantitative portfolio optimizations are very sensitive to the capital
market assumptions used; and
there are at least 44 such assumptions in the Asset-Liability Study, involving MPI’s 8-asset class
portfolio, as calculated below.
44 Important Assumptions (estimates, but “unknowns”)
8 average return assumptions (1 for each asset class)
8 volatility assumptions
28 correlations (= 8 x 7 ÷ 2)
44 “unknowns”
It goes without saying, perhaps, that an over-reliance on the quantitative results of optimizations is
inappropriate and imprudent, given the large number of “unknowns” and other considerations.
Public Plans? at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2070054.
P a g e | 42
Minimum/Maximum Constraints in Optimizations
The min/max constraints for asset classes in the Asset-Liability Study, listed below, are very binding. By
establishing them, the constrained optimization problem results in a “pre-determined” allocation to ≥
80% of the portfolio, leaving little room to optimize return/risk tradeoffs (i.e., allocating ≤ 20% of the
remaining assets).
80% of the portfolio is constrained31 as follows.
≥ 10% Canadian Equities
≥ 10% Real Estate
≥ 5% Infrastructure
≥ 55% Liability Matching
≥ 80% of the portfolio is “pre-determined” by minimum constraints
10% Minimum Weight to Canadian Equities to Retain a Meaningful Exposure to Home
Markets
MPI said that the minimum weight of 10% for Canadian equities was set “to retain a meaningful
exposure to our home equity markets”.32
It is not clear why it is important for MPI to maintain a “meaningful” exposure to Canadian equities
unless it relates to a regulatory requirement or return/risk consideration.
Pension plans and other institutional investors have generally acknowledged a “home bias” when it
comes to investing (i.e., overweighting domestic equity markets relative to the domestic market’s global
capitalization %), and they have been diversifying their portfolios internationally to remove the bias and
improve risk-adjusted returns.
This international diversification has been more important for Canadian investors, given the
concentrated nature of our market. Three (of 10) sectors in Canada, for example, typically represent a
very large proportion of the TSX market cap:



Financials > ⅓;
Energy ~ 20%; and
Materials ~10%.
Canadian equities are underrepresented in health care and information technology (< 3%), and Canada
represents a very small (~ 3%) portion of the world’s market capitalization of publicly-traded stocks.
31
Source: MPI’s response to CAC (MPI) 1-77
32
Source: MPI’s response to CAC (MPI) 2-39
P a g e | 43
Manitoba’s TRAF (Teachers’
Retirement Allowances Fund),
for example, has a higher
allocation of equities in
International Equities (33%)
than US Equities (30%) and a
combined foreign exposure
(63%) that is two times higher
than in Canada (37%) (far
left).33
TRAF’s allocation by sector
(near left) shows a more
diversified portfolio on a
sector basis as a result of
having more foreign
exposure.
When asked if MPI was comfortable with its concentration in Canadian Equities (⅔ of total equities),
given Canada’s small size (~ 3% of the world’s market cap) and concentrated exposure to three (of 10
GICS) sectors (Financials, Energy and Materials), MPI said it was “comfortable with the current exposure
to Canadian equities. The current portfolio was selected from the efficient frontier developed during the
Asset Liability Management (ALM) study completed in January 2015”.34
33
Source: TRAF’s annual report
34
Source: MPI’s response to CAC (MPI) 1-83
P a g e | 44
Equity Risk Concentration35
Most institutional investors have a significant
allocation to International Equities (i.e., excluding
Canada and US), but MPI has 0%.
Saskatchewan Auto Fund, administered by
Saskatchewan Government Insurance (“SGI”), for
example, has almost ¼ of its public equities in
International Equities as illustrated on the right.36
The average defined benefit pension plan, as
surveyed by the Pension Investment Association
of Canada (PIAC37), has an even higher allocation
to International Equities (> ⅓ of equites), with
non-Canadian Equities (US + International)
representing almost ¾ of total public equities and
Canadian Equities ~ ¼ as illustrated on the right.
Public Equity Mix
Canadian Equity
US Equity
International Equity
MPI
67
33
100
SGI Diff
47
20
31
2
23 - 23
100 -
Public Equity Mix
Canadian Equity
US Equity
International Equity
MPI
67
33
100
PIAC Diff
24
42
38 - 5
37 - 37
100 -
PIAC’s 2015 asset mix report is on the next page.
35
Source: CAC (MPI) 1-83
36
These figures were included in CAC’s questions and have not been updated since then.
Source: PIAC website, at http://www.piacweb.org/publications/asset-mix-report.html
37
P a g e | 45
P a g e | 46
Duration
MPI was asked if the duration-matching strategy would be as effective if future inflation (actual and/or
expected) turns out to differ from current expectations. MPI replied:
“MPI … agrees that the duration matching strategy is not as effective if future inflation turns out
to differ from current expectations … MPI “has accepted short term inflation risk and has
accounted for this risk through margins and reserve. The ‘excess portfolio’ was also designed to
provide some long-term protection against inflation.” The excess portfolio includes real estate
and infrastructure.”38
Borrowing (Interest Rate Risk) in Real Estate
The Real Estate Fund has leverage of ~ 30%, while Aon assumed an unlevered real estate investment
(i.e., zero debt), which means the gross and net (of debt) metrics differ between what was modeled by
Aon on the one hand in supporting the long-term asset mix decision and what was implemented in
practice on the other hand – a difference of 4% below.
MPI agreed that this implies a 4% difference between the actual exposures implemented by MPI and the
assumptions for “real estate” and “fixed income” in the Asset-Liability Study (i.e., gross real estate
exposure is closer to 13% than 9%, and exposure to fixed income is 4% lower – i.e., by the amount of
debt within the real estate portfolio). MPI added that it “reports the holding on a net basis in keeping
with the Net Asset Value reported by the fund manager. Therefore, it is not necessary to split out the
debt associated with the fund and report it separately”.39
When asked if the 4% difference (~ $95 million) was material, and whether the difference should be
reflected somehow for greater clarity (e.g., investment policy min/max constraints and targets) and
reflected in duration and/or other calculations, MPI said:
“The real estate pooled fund is reported net of debt as this represents the proceeds that would
be realized … upon liquidation ... Real estate investments are reported at their estimated fair
values as provided by the external manager in the audited financial statements for the fund.
The use of leverage in the real estate pooled fund represents a liability to that fund and is
integral to the manager’s strategy and cannot be separated. Because of these facts, there is no
need to consolidate the debt associated with the investment in GREF with the Corporation’s
fixed income portfolio and therefore no impact on duration.
38
39
Source: MPI’s response to CAC (MPI) 76
Source: CAC (MPI) 76
P a g e | 47
When the real estate pooled fund is measured on a net basis, the overall exposure to real estate
(pooled real estate fund and CityPlace) as of February 2016 was 11.3%, which is within the
min/max range defined for that asset class (7% - 13%). The rebalancing policy defines the
monitoring of the target allocation on a fiscal quarterly basis, at a minimum, and rebalancing
back to target within six months, if the weight of any asset class falls outside the allowable
range. Because the real estate pooled fund is reported on a net basis there is no need for
additional clarification of the investment policy min/max constraints and targets.”
For greater clarification, here are a few observations.



There is a 4% difference between what was modeled by Aon (and presumably the basis for the
policy allocations implemented) and what was actually implemented in practice:
o Real estate’s gross exposure (excluding debt) is 4% higher than Aon’s recommendation;
o Net exposure to interest rates, at the total portfolio level, is 4% lower even though the
effect of leverage is accounted for within the real estate asset class;
MPI did not answer our question regarding the materiality of the 4% difference, where (though
not stated) implicitly materiality refers to the effect that the 4% difference might have on
portfolio decision-making or design as distinct from how items might be accounted for;
o Given MPI’s duration policy and interest rate risk management practices, which are very
detailed, the risk management (as distinct from the accounting) treatment for liabilities
arising from real estate activities is inconsistent with those for the liabilities arising from
providing auto insurance (and pensions);
I disagree with MPI’s answer that “because the real estate pooled fund is reported on a net
basis there is no need for additional clarification of the investment policy min/max constraints
and targets”. I would note that the policy min/max constraints and targets approved are
therefore not consistent with the results and recommendations from the Asset-Liability Study
o Greater clarity and consistency of treatment between the Asset-Liability Study and the
policies would seem both more appropriate and prudent.
Commingling of Employee Benefit Plans
When asked if the creation of a separate trust for the pension plan would “trigger” a realization
(recognition in net income) of “remeasurement of employee future benefits”, MPI said:
“This question is theoretical in nature and therefore not relevant to the rates setting process.
Further, without discussion on the nature of the trust and the type of transaction to create and
fund the trust, an accurate answer can not be provided.”40
I believe strongly that the portfolio design of the whole portfolio (including the pension plan) should not
be influenced by how the parts (insurance vs pension liabilities) are accounted for.
40
Source: CAC (MPI) 2-34
P a g e | 48
APPENDIX A: VALTER VIOLA’S CURRICULUM VITAE
VALTER VIOLA
50 De Vere Gardens, Toronto, ON M5M 3E7 | [email protected] 416 819 2307
PROFILE
Portfolio, research and risk management professional with over 20 years of experience in institutional
investment management (mostly defined benefit pension plans)


Executive and senior management roles in investment research, economics and risk management at
two of Canada’s largest institutional investors
Consultant to large North American institutional investors, advising Boards, Investment Committees
and Management Teams on investment strategies and investment risk management
PROFESSIONAL EXPERIENCE
2014 – 2016
MaPLE Toronto, Santiago
Partner

2005 – 2014
Participating in private energy and infrastructure opportunities in South America,
starting with power generation and related infrastructure in Chile
Holland Park Toronto
President, Founder


Advised defined benefit pension plans, workers’ compensation funds and other
institutional investors on investment risk governance, management and
measurement practices, including:
o Board education;
o developing investment/risk frameworks;
o drafting investment/risk management policies and procedures that were
prudent and appropriate in the circumstances and took into account
relevant principles and leading practices; and
o developing risk budgets to support the management of surplus (assets and
liabilities) and active management programs (performance vs benchmarks)
Provided investment risk monitoring and reporting services to pension funds,
including:
o managing third party risk analytics, clients’ holdings and benchmark data,
proxies for liabilities and market data;
o recommending and implementing generally accepted and appropriate risk
measurement methodologies; and
o preparing and presenting reports to Boards and executive management
teams to meet their needs for relevant, reliable and timely information
about investment risks
P a g e | 49
VALTER VIOLA
2000 – 2005
CPP Investment Board Toronto
Vice President, Research and Risk Management




1993 – 2000
First executive responsible for total portfolio research, design and investment risk
management of the largest single purpose pool of capital in Canada
Led a growing team of professionals, focused on the total portfolio including:
o investment risk management (relative to liabilities and benchmarks);
o policy asset mix and currency hedging;
o active management;
o other investment policies
Collaborated with the CEO/CIO, VP Private Markets and VP Public Markets in the
development and implementation of investment strategies
Collaborated with other executives to develop and implement strategies and
business plans, policies and procedures, including leading the development of an
investment/risk management framework that took into account the unique
circumstances of the CPP and CPPIB (e.g., large unfunded liability, non-marketable
bonds, large cash inflows)
Ontario Teachers’ Pension Plan Toronto
Director, Portfolio Manager, Analyst (Research and Economics)




1992 – 1993
Member of the Investment Planning Committee, with shared responsibilities to
advise the CIO on the tactical management of the total portfolio (shorter-term
horizon, broad asset class allocations and currency hedging decisions)
Supported strategic/policy and tactical asset mix/currency hedging and other total
fund decisions through independent research, including:
o developing the fund’s first asset/liability model, which supported the fund’s
asset mix transition shortly after the fund’s inception;
o conducting research to support new asset class introductions; and
o recommending appropriate benchmarks
Managed the tactical asset allocation portfolio, a portfolio that had one of the largest
value added targets for the fund
Managed the real return bond portfolio, including closing the largest single
investment in the fund’s history ($650 million private placement of inflation-linked
bonds that financed the 407 Electronic Toll Road)
Wilfrid Laurier University and York University
Lecturer in Investments, Finance and Accounting

1990 – 1992
Conducted lectures in the undergraduate programs at two universities
Corporate Planning Associates
Financial Advisor

1986 – 1988
Advised high net worth individuals (portfolio management, tax planning, etc.)
Price Waterhouse
Auditor

Completed audit and related work for corporate clients in various industries, as well
as organizations in non-profit and other sectors
P a g e | 50
VALTER VIOLA
OTHER EXPERIENCES
PENSION ASSOCIATIONS/COMMITTEES
2006 – 2009 Healthcare of Ontario Pension Plan (HOOPP)
External Advisor to Investment Committee

Advised the Investment Committee of a large, Canadian defined benefit plan on
matters related to the management of the total portfolio
2003 – 2005 Pension Investment Association of Canada (PIAC)
Member of Investment Practices Committee


Shared non-proprietary investment practices with peers as a member of an industry
association
Led the publication of a paper (“Risk Budgeting”) to meet the needs of member
organizations
VOLUNTEER
2013 – 2016 Enbridge Ride to Conquer Cancer
Fundraiser, Co-Captain
2004 – 2007 North Toronto Soccer
Coach
1988 – 1989 New Visions Toronto
Board Member
EDUCATION
1995
Chartered Financial Analyst
1990
Master of Business Administration, University of Western Ontario
1989
Chartered Accountant
1986
Bachelor of Commerce, University of Toronto
P a g e | 51
APPENDIX B: GLOSSARY OF TERMS
The table below includes the definitions of commonly used terms.
Concept
Definition
Accounting reports
Reports prepared by accountants to meet specific needs (which may differ
from the needs of users of other types of reports).
The basis of measurement in MPI’s accounting reports is “historical cost …
except for financial instruments and insurance contract liabilities and
reinsurers’ share of unpaid claims ... Insurance contract liabilities … are
measured on a discounted basis in accordance with accepted actuarial practice
(which in the absence of an active market provides a reasonable proxy for fair
value) ...”
Actuarial reports
Reports prepared by actuaries to meet specific needs (which may differ from
the needs of users of other types of reports).
Asset risk
Market risk related to assets measured in an absolute sense (i.e., not relative
to a benchmark or liabilities).
Assumptions
The inputs used in a valuation.
Available for sale
(AFS)
As defined in MPI’s Financial Statements, AFS assets represent ~ 20% and
include:



Equity Investments
Cash and Cash Equivalents
Other Investments (not material).
Duration
Market risk metric that measures the price sensitivity of a security or portfolio
to changes in interest rates.
Financial assets and
liabilities at fair value
through profit or loss
(FVTPL)
As defined in MPI’s Financial Statements, FVTPL assets represent ~ 54% and
include:
Held to maturity
(HTM)
As defined in MPI’s Financial Statements, HTM assets represent ~ 25% and
include:






Other Bonds (Federal, Other Manitoba, Other Provinces, Corporates)
Infrastructure
Pooled Real Estate Fund.
Bonds – Manitoba
Municipal
Bonds – Manitoba, Schools.
P a g e | 52
Concept
Definition
Inherent risk
The “true” market risk of an asset or liability.
Investment risk
Risk of loss associated with investment activities, including:


market risk (including credit risk)
liquidity risk.
Liability from Basic
Claims
Liability or provision at a point in time arising from providing basic auto
insurance coverage.
Liability from pension
plan
Liability or provision at a point in time arising from providing defined benefit
pension coverage to employees.
Liquidity risk
Risk of not being able to generate sufficient cash or its equivalent to meet
commitments as they come due in a way that is:


timely and
cost effective.
Market interest rate
A rate used to value an asset or liability based on an assessment of the
inherent market risk of a security and/or its underlying cash flows
Market reports
Reports prepared to support market risk management, including portfolio
design (e.g., asset mix).
Reports are prepared using “best estimates” (i.e., no bias towards
conservatism).
Market risk
Risk of loss in the market value of assets and/or liabilities due to changes in
security prices, interest rates, currencies or other such factors in either
absolute (e.g., assets only) or relative terms (e.g., vs benchmarks or liabilities,
known as active risk/tracking error and surplus risk respectively).
Other comprehensive
income (OCI)
As defined in MPI’s Financial Statements.
Reinvestment rate
risk
Risk that reinvested income will be insufficient to meet a goal.
Risk
An outcome which has some (non-zero) probability of having an adverse
impact on one or more stakeholders.
Risk (as defined in
Asset-Liability Study)
An important consideration in portfolio design (i.e., asset mix).
Risk-free rate
A term used by MPI in response to a Round 1 question; not a term to be used
in responses to questions below, unless “risk-free rate” is clearly defined by
MPI and distinguished from other “rates”.
P a g e | 53
Concept
Definition
Risk profile
Risk level and/or risk allocation (composition).
Surplus risk
Market risk related to the difference between assets and liabilities.
Unrealized gains and
losses
Gains and losses, representing differences between the cost (accounting or
book value) of one or more assets and their fair value for accounting purposes.
Value
The quantitative assessment of an asset or liability based on the application of
a valuation methodology and related assumptions (e.g., book value, accounting
value, actuarial value, market value).
Valuation
methodology
The method, excluding the assumptions used to implement the method, to
value assets or liabilities for one or more purposes (e.g., discounted cash flow
or present value methodology).